|
Title: Proof of Transactions Post by: coradan on January 05, 2014, 06:26:10 PM Hey guys,
Yesterday talking with Zesterer in fastcointalk.org (http://www.fastcointalk.org) forums, he told to me that he with a friend talked about to make a some coin that rewards its use. Thinking on this idea, the system Proof of Transactions (PoT) came to my mind... It is a system where the blocks are discovered by who has more transactions. It is simple: Hashing would be fixed to one hash per second for everyone. The difficulty would not be equal for everyone, it would be variable and less for addresses with more transactions. And when somebody discovers one block, his transactions count would be reset to 0, and the count would start again until the next discovery. The reward would be the fees of transactions in the block. It is very simple, but I think that system promotes the use of the coin, as Zesterer said. And avoids 51% attack because in each discovery resets the discovery probability to 0 to the account did it. I think it would be a good system to implement a future version of a coin to promote the use and distribution more equity and faster of the coin of our day to day... :) It does not reward the huge amount of money, neither who buy more technology and spend more energy... simply rewards to people spend their money. This coin would not have mining, all amonut would be generated at the begining, so it would be necessary a good distribution system (equity and fair)... It is a today idea, I dont have details about it yet... so I hope your opinions and pros and cons that you see it... Thanks in advanced...! Title: Re: Proof of Transactions Post by: coradan on January 05, 2014, 06:46:08 PM Perhap better name would be Proof of Use (PoU)...
Title: Re: Proof of Transactions Post by: markm on January 05, 2014, 08:52:39 PM So what kind of passports or drivers license or whatever do you plan to use to try to not count people who have thousands of wallets that send "transactions" back and forth between them just to create fake "activity" ?
-MarkM- Title: Re: Proof of Transactions Post by: tk808 on January 05, 2014, 08:54:32 PM This is what Aerocoin is claiming to do basically. You get Aeros through transactions via interest.
But i think all those IPO coins are scams Title: Re: Proof of Transactions Post by: black_swan on January 05, 2014, 08:58:28 PM So what kind of passports or drivers license or whatever do you plan to use to try to not count people who have thousands of wallets that send "transactions" back and forth between them just to create fake "activity" ? -MarkM- That's exactly what I thought, 1) cheating is too easy. 2) How do you spend a new coin? Nobody is going to accept it 3) PoS: as you said, find a distribution method is a hell of a nightmare, it will never be equal Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 05:10:10 AM With respect point one:
I think fees would avoid this, because if you make a transactoin, you will pay fees and this fees can get back to you if you confirm the transaction, but other can confirm it then you will loose fees, then would not be profitable to do transactions between oneself... I do not understand this point: 2) How do you spend a new coin? Nobody is going to accept it why not? Only the input address resets to 0 its transactions accumulated count. And about distribution problem, I think here there is no solution yet... Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 07:37:03 AM Imagine a person is doing transactions between two address to increase his transactions counter to obtain high probability to get a block, when he achieves the block, his transactions counter resets to 0, and until this moment he has spent a large amount of his coins in fees to increase his transactions counter. And his reward can be very little, because imagine that he discovers a block with a total fee of 1 coin...
He wont try the fake activity again! :) Title: Re: Proof of Transactions Post by: BitThink on January 06, 2014, 07:41:43 AM Hey guys, How to fix everyone's hashing rate to one hash per second? Relying on no one changes your source code, or is there any method to ensure this in the decentralized network?Yesterday talking with Zesterer in fastcointalk.org (http://www.fastcointalk.org) forums, he told to me that he with a friend talked about to make a some coin that rewards its use. Thinking on this idea, the system Proof of Transactions (PoT) came to my mind... It is a system where the blocks are discovered by who has more transactions. It is simple: Hashing would be fixed to one hash per second for everyone. The difficulty would not be equal for everyone, it would be variable and less for accounts with more transactions. And when somebody discovers one block, his transactions count would be reset to 0, and the count would start again until the next discovery. The reward would be the fees of transactions in the block. It is very simple, but I think that system promotes the use of the coin, as Zesterer said. And avoids 51% attack because in each discovery resets the discovery probability to 0 to the account did it. I think it would be a good system to implement a future version of a coin to promote the use and distribution more equity and faster of the coin of our day to day... :) It does not reward the huge amount of money, neither who buy more technology and spend more energy... simply rewards the people spends his money. This coin would not have mining, all amonut would be generated at the begining, so it would be necessary a good distribution system (equity and fair)... It is a today idea, I dont have details about it yet... so I hope your opinions and pros and cons that you see it... Thanks in advanced...! Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 07:43:37 AM How to fix everyone's hashing rate to one hash per second? Relying on no one changes your source code, or is there any method to ensure this in the decentralized network? For example, PPC already do it. And NXT I think too... Every PoS coins do it. A kind of timestamp control... Title: Re: Proof of Transactions Post by: BitThink on January 06, 2014, 07:54:26 AM How to fix everyone's hashing rate to one hash per second? Relying on no one changes your source code, or is there any method to ensure this in the decentralized network? For example, PPC already do it. And NXT I think too... Every PoS coins do it. A kind of timestamp control... Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 08:06:52 AM In PoS making more hashes, you would obtain more benefits too... It has a kind of timestamp control to avoid this.
PoS do two things: 1.- Fixed hash to one hash per second. 2.- Difficulty is not fixed, it is variable and less for the addresses with more number of coin-age (coins*time_in_wallet). PoT would do two things too: 1.- Fixed hash to one hash per second. 2.- Difficulty is not fixed, it is variable and less for the addresses with more number of accumulated transactions. Title: Re: Proof of Transactions Post by: BitThink on January 06, 2014, 08:20:51 AM In PoS making more hashes, you would obtain more benefits too... It has a kind of timestamp control to avoid this. I am not sure how PoS in PPC works, but in nxt, they claim that the mining is deterministic and who should forge a block is determined as long as that node is forging. PoS do two things: 1.- Fixed hash to one hash per second. 2.- Difficulty is not fixed, it is variable and less for the accounts with more number of coin-age (coins*time_in_wallet). PoT would do two things too: 1.- Fixed hash to one hash per second. 2.- Difficulty is not fixed, it is variable and less for the accounts with more number of accumulated transactions. I am quite curious about how timestamp control can avoid a client to make multiple hash in a second. It may avoid a client to SUBMIT more than a hash a second, but hashing is done locally and no other node will know how many times I have tried. Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 08:25:04 AM Yep, in NXT with transparent mining fix the node, I dont know how do it...
But fix the hash to one per second is possible. Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 08:43:08 AM I am quite curious about how timestamp control can avoid a client to make multiple hash in a second. It may avoid a client to SUBMIT more than a hash a second, but hashing is done locally and no other node will know how many times I have tried. I can see that it is simple: in PPC the Nonce is a Timestamp of one second. Title: Re: Proof of Transactions Post by: BitThink on January 06, 2014, 09:13:29 AM I am quite curious about how timestamp control can avoid a client to make multiple hash in a second. It may avoid a client to SUBMIT more than a hash a second, but hashing is done locally and no other node will know how many times I have tried. In PPC the Nonce is a Timestamp of one second. Thanks for your information. That make sense to ensure there's only one hash in one second. I still have a doubt on this, though. It's possible that the nonce of current timestamp does not lead to a solution for all clients, so the current second has to be skipped and the new nonce of next second has to be issued. Therefore, a clever client may predict this possibility and simultaneously hash with the current timestamp, next timestamp, next next timestamp, ..., as the nonce. That being said, this does not bring too much advantage to the fast computers, because there're at most couple of tens nonce to try before a solution is found anyway. I admit that control the hashing rate in this way is possible now. Ok, then back to PoT vs PoS. Stake is more difficult to be faked than transaction numbers. Fees may help preventing average user to take advantage of high transaction numbers, but may be enough to stop 51% attacking. Moreover, usually the number of normal transactions of a new coin is very low, and there's not too much cost for a person to fake transactions to get the advantage in mining. If more and more people doing this, the system will be flooded with meaningless transactions and becomes both slow and bloated. Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 09:49:09 AM Ok, then back to PoT vs PoS. Stake is more difficult to be faked than transaction numbers. Fees may help preventing average user to take advantage of high transaction numbers, but may be enough to stop 51% attacking. Moreover, usually the number of normal transactions of a new coin is very low, and there's not too much cost for a person to fake transactions to get the advantage in mining. If more and more people doing this, the system will be flooded with meaningless transactions and becomes both slow and bloated. I am not agree with this, because if you try to fake transaction numbers this has a cost for you and when you reach to the desired reward your transactions counter will reset to 0. And your desired reward can be very little, it depends of fees on the block you had discovered. Then you has spent large money to obtain a privileged position in transaction number counter to nothing. With respect 51% attack, suppose that you have one address that achieves the 51% of transactions numbers, then you discover a block and your transactions counter go back to 0. Your 51% has been useful to one block only. If you want to continue with the attack, you will have to obtain the 51% again, and spent money again in fake transactions with their fees to achieve it... The system self protects itself. With this system the number of normal transactions would be high, because they are rewarded... Title: Re: Proof of Transactions Post by: CaptainBeck on January 06, 2014, 09:51:58 AM Surely the exchanges will always win the blocks because they have most transactions out of all the wallets??
Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 09:59:14 AM Surely the exchanges will always win the blocks because they have most transactions out of all the wallets?? It could be, only the withdrawals would be transactions that increase the transactions counter of the addresses of the exchanges. yep, this can see as critical point at the current moment when the coins are using for trading only... But in the future the exchanges will have less relevance. And this system promote the use of the coins, so perhaps it would accelerate the time of the adoption of crypto-coins in our day to day... The current situation with the exchanges could see as a support for the security of the system too. And they can reward the withdrawals like if they were pools... :) Title: Re: Proof of Transactions Post by: marakanito on January 06, 2014, 10:15:53 AM Proof of Transaction = Dices would mine a lot.
What about proof of transaction*amount (or fee)? Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 11:55:33 AM Proof of Transaction = Dices would mine a lot. What about proof of transaction*amount (or fee)? it is possible... But in this way you reward the huge amounts of money and it would be like PoS. Title: Re: Proof of Transactions Post by: Anillos2 on January 06, 2014, 03:30:15 PM What happens if you add dummy transactions? These transactions could act like a second nonce. Also, You can change the order of transactions, so You Could have more than one hash per second.
Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 03:53:35 PM Studying the issue in more depth I see a possible problem: it would be a parallel blockchain hidden to the network with only two addresses that could generate a fake blockchain with more transactions than the real blockchain...
I will think if is possible solve this issue... Title: Re: Proof of Transactions Post by: Bansheroom on January 06, 2014, 04:00:05 PM To avoid abuse you could implement a system that uses lower fees (perhaps higher PoT-Paiout), the longer the coins have not been moved before.
Title: Re: Proof of Transactions Post by: coradan on January 06, 2014, 04:35:33 PM To avoid abuse you could implement a system that uses lower fees (perhaps higher PoT-Paiout), the longer the coins have not been moved before. What do you mean with lower fees? |